RFR(XS): 8202425: [s390] C2: Wrong unsigned comparison with 0

Doerr, Martin martin.doerr at sap.com
Mon Apr 30 15:01:22 UTC 2018


Hi Aleksey,

unfortunately, we're using "branch if >".
"branch if !=" would be equivalent and not cause this error.

Removing the node only leads to larger instructions getting emitted which should be affordable.

Best regards,
Martin


-----Original Message-----
From: Aleksey Shipilev [mailto:shade at redhat.com] 
Sent: Montag, 30. April 2018 16:52
To: Doerr, Martin <martin.doerr at sap.com>; 'hotspot-compiler-dev at openjdk.java.net' <hotspot-compiler-dev at openjdk.java.net>
Subject: Re: RFR(XS): 8202425: [s390] C2: Wrong unsigned comparison with 0

On 04/30/2018 04:18 PM, Doerr, Martin wrote:
> compU_reg_imm0. It's used for unsigned comparison with 0, but it's implemented as signed
> comparison.

I am curious. Why do we care about signness when comparing to zero?

> http://cr.openjdk.java.net/~mdoerr/8202425_s390_cmpU0/webrev.00/

Looks good.

-Aleksey



More information about the hotspot-compiler-dev mailing list